- 期刊
Research on Face Recognition Method Based on Deep Neural Network
摘要
In this paper, a scheme based on the improved network structure model is proposed to solve the problems of slow loss gradient descent, slow recognition rate gradient rise and poor stability in face recognition of AlexNet network.Based on AlexNet network, the first two convolutional layers are decomposed and filled, while one convolutional layer and one full connection layer are removed.The structure consists of 6 convolutional layers, 3 pooling layers and 2 full connection layers.We compared the performances by testing face recognition on ORL face dataset and found that the improved convolutional kernel network is superior to AlexNet network in general.
延伸閱讀
- 林仁信（2018）。Face Recognition Algorithm Based on Deep Learning on Desktop Device System〔碩士論文，國立暨南國際大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0020-1901201913392500
- Kao, H. P. (2018). Face Recognition Using Small Amount of Data with Deep Learning [master's thesis, National Taiwan University]. Airiti Library. https://doi.org/10.6342/NTU201801859
- Li, Z. (2020). Research on Face Detection Based on Deep Learning. World Scientific Research Journal, 6(3), 87-95. https://doi.org/10.6911/WSRJ.202003_6(3).0012
- 蔡佳韋（2021）。Face Recognition Based on Relation Neural Networks and Embedded Systems〔碩士論文，國立臺灣師範大學〕。華藝線上圖書館。https://doi.org/10.6345/NTNU202101301
- 陳郁萍（2022）。Research on Face Recognition Using Convolutional Neural Networks〔碩士論文，國立宜蘭大學〕。華藝線上圖書館。https://doi.org/10.6820/niu202200093